JPS6184550A - Inspecting device using nuclear magnetic resonance - Google Patents

Abstract

PURPOSE:To correct the influence of the irregularity of a static magnetic field and/or nonlinearity of a slanting magnetic field by measuring the intensity distribution in the visual field of the static magnetic field and/or slanting magnetic field, and correcting an image obtained in the presence of the magnetic field at each point. CONSTITUTION:A computer 1 outputs various instructions to respective devices at constant timing. The output of a high frequency pulse generator 2 is amplified by a power amplifier 3 to establish a high frequency magnetic field and also excite simultaneously a coil 4 which detects a signal generated by an object body 16. The coil 4 serves as a receiving coil, and a received signal component is passed through an amplifier 5 and detected by a detector 6, and then inputted to and processed by the computer 1, so that the signal is displayed on a display device 7 after being converted into an image. Coils 8, 9, and 10 generate slanting magnetic fields in the z direction and directions perpendicular to it. A static magnetic field is produced by a coil 14 which is driven by a power source 15. A human body 16 to be inspected is laid on a bed 17, which is movable on a support table 18.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to an inspection device using nuclear magnetic resonance (hereinafter referred to as rNMRJ), particularly in NMR imaging.
The present invention relates to an inspection apparatus using NMR that can completely and independently correct the effects of static magnetic field inhomogeneity and gradient magnetic field nonlinearity that cause image quality deterioration.

[Background of the invention]

2. Description of the Related Art Conventionally, X-ray CT and ultrasonic imaging devices have been widely used as devices for non-destructively inspecting internal structures such as the head and abdomen of a human body. In recent years, attempts to perform similar tests using nuclear magnetic resonance phenomena have been successful, and it has become clear that information that cannot be obtained with X1CT or ultrasound imaging devices can be obtained. In an inspection device that uses nuclear magnetic resonance phenomena, it is necessary to separate and identify signals from an inspection object in correspondence with each part of the object. One method is to obtain position information by applying a gradient magnetic field to the object to be inspected, varying the static magnetic field placed on each part of the object, and thereby varying the resonant frequency or phase encoding amount of each part. .

Regarding its basic principle, Kumar et al.
gn.

Re5on, (Yo 8, 69 (1975)) or Edelstein et al.
BioL, (751 (1980)), so it will be omitted here.

In such imaging, non-uniformity of the static magnetic field and non-linearity of the gradient magnetic field are causes of deterioration of image quality. It has been reported that the effects of static magnetic field inhomogeneity and gradient magnetic field nonlinearity can be corrected if their distributions are known (Se
kihara at al.

Phys, Mad, Biol, Dan, l 5
(1984)).

A method for measuring the non-uniform distribution of static magnetic fields among these distributions has already been reported by Maudsley et al. in J. Phys.
.

E: Sci, Instrument, Rinishi 216 (1
As reported in 984), it is possible to obtain even the distribution of picture elements. However, until now, there was no known method for determining the nonlinearity of gradient magnetic fields. Since the coil that generates the gradient magnetic field is usually installed inside the bore of the magnet for generating the static magnetic field, it is difficult to improve its linearity beyond the current level, and depending on the object to be measured, the effect of this cannot be ignored. Therefore, it has not been possible to perform sufficient correction only by correcting the static magnetic field.

[Purpose of the invention]

The present invention was developed in view of these drawbacks, and its purpose is to detect static magnetic field inhomogeneity and gradient magnetic field nonlinearity separately, and to make it possible to correct the influence of at least one of them. The purpose is to provide an inspection device that

[Summary of the invention]

The gist of the present invention is that in an inspection device using the Fourier imaging method, two types of gradient magnetic fields with different strengths (assuming the distributions are the same) are used to determine the non-uniform distribution of static magnetic fields and the distribution of gradient magnetic fields within the field of view. The main feature of this method is that the images are corrected based on the detected data.

Regarding this, some supplementary explanation will be given below.

To simplify the explanation, a two-dimensional modified spin warp method, which will be described later, will be explained as an example. .

Let ρ(x, y) be the target spin density distribution, E(x r
y) is the deviation of the static magnetic field distribution from the reference amount, G, , G,
Let be the magnitude of the gradient magnetic field in the X direction and the X direction, respectively.

Here, the nonlinear gradient magnetic field in the Xy and X directions is G, (x
tt, (x, y)) t ay (y+εy(xt
It is expressed as y)). Expressed in this way, G yo, G is proportional to the current flowing through the coil, and is 5w (xt y) + εy (X
ty) is a quantity determined by the shape of the coil. here,
εm(Xly) r εy(xty) represents a component that deviates from those straight lines. At this time, the measured two-dimensional signal s(a-+ty) is 5(a-+ty)=z7p(xty)exptoj y
(t* (xt 5, (x, y)
(xty) + Gy r, y (x, y) ÷ Gyy) ty)
) d-dy (1). γ is the nuclear gyromagnetic ratio. Here, if the integral variable is -S (G, yty)=, //'p (x'*y'
) exp(-jy (t, x' G.

+tyy' ay)) d-'d,' (3) Here,
The relationship expressed by the following equation holds true between ρ(Xly)=ρ'(x'Ty').

ρ'(x' + y') = ρ(X, y)/J(x, y
) (4) However, J(x, y) is given by the following formula.

Equation (,7) is Eyo, when the change in E is small, a,
It can be expressed as Gyay.

By the way, as can be seen from equation (2), the distribution ρ(x, y) of the target object is affected by the inhomogeneity of the static magnetic field -E(x, y) and the nonlinearity of the gradient magnetic field @, (xt y) and εy ( xty) changes to ρ/(x/, yl), it is not possible to separate the effects of both.

However, it has been found that by using a plurality of gradient magnetic fields having different magnitudes as G, it is possible to separate the effects of both. As an example, suppose two types of magnetic fields are used, and the magnetic fields are G, l, G, .

As a formula corresponding to formula (2), is obtained.

The following equation is obtained from equations (8) and (9).

What this equation means is the distortion of the image obtained for G2 and G. This means that only the inhomogeneity -E(xvy) of the static magnetic field can be extracted purely from the difference between the distortion of the image obtained and the distortion of the image. Note that since the field of view of the image obtained is different between G,1 and G,2, the time interval for sampling the signal is
, □ and G F2 so that they have the same field of view, or it is necessary to make the field of view the same by enlarging or contracting the obtained image.

Once E(Xly) is found, E y (X*y) can be found from equation (8) or equation (9). On the other hand, E-(xs
y) can be found from x/-X or x'-x.

In the above explanation, X and x/(x#) and y and y'
It is assumed that a correspondence relationship with (y') is required. To find these relationships, 1. For example, use a phantom in which rod-shaped rice plants 21 having a predetermined spin density as shown in FIG. All you have to do is find the positional relationship between the corresponding points of the image obtained in 2) and G, 2. The circle mark in the figure indicates the position where the sample is placed6. It is also possible that the sample is missing only in the marked area and the sample is clogged in other areas. In any case, any phantom that can detect image distortion will suffice. Now ρ(X,y)
If we know that corresponds to the image ρ′(y′,y′), then
The correspondence relationship between and y' and y and y' is required. If we find this for all x and y, we will find all the correspondence relationships in the field of view6x and x #
, y and y', ρ'(X'ry') can be found in exactly the same way, with G being Go.

[Embodiments of the invention]

Embodiments of the present invention will be described in detail below with reference to the drawings.

First, let's take the case of imaging a two-dimensional surface as an example.
The principle of the modified spin warp method and an example in which the present invention is applied to the two-dimensional modified spin warp method will be described. Figure 3 shows the irradiation pulse for implementing the two-dimensional deformed spin warp method,
It shows the timing of signals from the gradient magnetic field in the XtV direction and the nuclear spin. Here, a certain cross section parallel to the (X, y) plane is selected. In the figure, RF indicates the above-mentioned irradiation pulse, and G and Gyo indicate gradient magnetic fields in the y and x directions, respectively. Further, S indicates a signal from nuclear spin.

First, a 90'' RF pulse is irradiated to tilt the nuclear spin in the sample by 90''. Immediately thereafter, the gradient magnetic field G is applied for a time t,
180' RF pulse is applied fifth time. The i-measurement of the signal is performed while applying G.

The two-dimensional signal S(G*+ ty) obtained as a result of performing such measurements by changing the magnitude of the gradient magnetic field in the X direction
is between the nuclear spin distribution ρ(X, y) of the selected cross section and S (G, +ty) = ff p (x+y)exp
(-j y (G, xt.

+ G F yty)) There is the relationship shown in (11). However, equation (11) only holds true when the slope of G, , G is linear and the static magnetic field is uniform. As can be seen from equation (11), the nuclear spin distribution ρ(x, y) of the selected cross section is found by performing a two-dimensional inverse Fourier transform on S (G −+ t y ). The above is the principle of the modified spin warp method.

Now, if the static magnetic field is not uniform within the field of view and the gradient magnetic field has nonlinearity, then S(G,

ty) and ρ(xey) is expressed not by equation (11) but by equation (1) above. Then, the measurement signal S(G,.

t,)i The image obtained by two-dimensional inverse Fourier transformation is ρ
'(x'ry' ), which is the original distribution ρ
The image density corresponding to the nuclear spin distribution ρ(x, y) at the point (X, y), which is equal to (:cty) subjected to the coordinate transformation shown by equation (2), is ( 2) It will be reproduced at the point (x' * y''') shown by the equation.

The present invention takes advantage of this fact and calculates ρ'(x' + y'
) to obtain an image corresponding to the original distribution ρ(x, y), the gradient magnetic field G□( Xl yLGyx
(x+y), the inhomogeneity of the static magnetic field -E(xty) and the nonlinearity of the gradient magnetic field εjX+3'L εy(:ct y)
is measured separately, and ρ(x+y) is obtained based on the data using equation (8) or equation (9).

By the way, in reality, ρ(X, y) is calculated at discrete points, so if one image is calculated using an NXN matrix, ρC
x, y) as ρ(I, J) (1=0.1°...N-1:
J=0,1. ...N-1).

In addition, the deviation of the static magnetic field on the pixel point (I, J) from the standard value, the deviation of the gradient magnetic field from the straight line f*(Xl yLεycx
+ y) are also E(IT JL t, (x, JLε
,C1,J) written as <,a,1, the reproduced image data ρ
'(X'ty') is also found at discrete points, so ρ'C
I', J'). Furthermore, since the reproduced image data ρ'(x' * y') at ayz is also determined by discrete points, it is expressed as ρ'(I',J').

Now, the value ρ(I, J) at one pixel point (I, J) after correcting the static magnetic field inhomogeneity at one pixel point (I, J) and the nonlinearity of the gradient magnetic field is calculated in the reproduced image by However, the point (ξ, η) is generally the pixel point (I', J') (I'0.1...
N-1: J' = 0.1. ...N-1). Therefore, for example, the value of ρ'(ξ', η') is set to (ξ'
, η') may be considered. In other words, let. Here, the symbol [ ] shall mean the largest integer not exceeding the value in [ ]. Using the result of equation (13), g = (1-A) (1-A2) ρ' (11j) + (
1-Δt) IJzρ'(1+j+1)+A(1-A
1) ρ'(11,]+1)+Δ□Δ2ρ' (4+ 1
, j+1 ) (14).

ρ(I, J)=g (15
). When ρ(I, J) ignores interpolation errors, an image is obtained in which the inhomogeneity of the static magnetic field and the nonlinearity of the gradient magnetic field are corrected.

FIG. 4 shows a configuration diagram of an inspection device that is an embodiment of the present invention. In the figure, 1 is a computer, 2 is a high-frequency pulse generator, 3 is a power amplifier, 4 is a coil for generating a high-frequency magnetic field and at the same time detecting the signal generated from the target object 16, 5 is an amplifier, 6 is a detector, 7 is a signal processing device.

Further, 8, 9 and 10 are coils that generate gradient magnetic fields in two directions and in a direction perpendicular to these, respectively; 11, 12;
13 is a power supply unit that drives the coils 8, 9, and 10, respectively.

The computer 1 also has a function of outputting various instructions to each device at a constant timing. The output of the high frequency pulse generator 2 is amplified by a power amplifier 3 to excite the coil 4. The coil 4 also serves as a receiving coil as described above, and the received signal component passes through the amplifier 5 and is detected by the detector 6, then input to the computer 1, where it is converted into an image on the display 7 after signal processing.

Note that the static magnetic field is generated by a coil 14 driven by a power source 15. The human body 16, which is the object to be inspected, is placed on the bed 17.
The bed 17 is configured to be movable on a support stand 18.

In addition, 19.20 is a storage device (hereinafter referred to as "memory")
The memory 19 stores the image ρ'(I'
, J' L ρ'(X'ry' ) are stored in the memory 20, and (i, j) and corresponding to all pixel points (I, J) calculated from equation (13) (
Δ4. Δ2) is stored. This (i, j).

Equation (8) is used to calculate (Δ0.Δ2). In the inspection apparatus configured as described above, the computer 1 calculates (i, j) and (IU, Δ) for the corrected pixel point (I, J).
2) from the memory 18, i+1. j+1
．． From the primary memory 19 to calculate 1-Δ1, 1-72, ρ/ (i/ , j7 ), ρ'(i+1゜j'), ρ
'(1' z j+1) and ρ' (i+1゜j+
1), calculate equation (14), and use the result as the interpolated image ρ(I, J) at the pixel point (L J). J) and display the results on the display 7.

By the way, when calculating equation (13), E, C
I, J), ε, (I, J), and E(I, J) must already be found. To obtain these, the phantom shown in FIG. 1 is used as described above. In this phantom, the sample is placed at the position indicated by the circle, or the sample is removed only from that part. The following explanation is for convenience only.

We will focus on the former case. Assume that the image obtained when a gradient magnetic field of G, 1 is applied is shown in FIG. If it is known that the image 22 corresponds to the sample 21, the coordinates Cx, y) of the sample 21 and the coordinates Cx' + y of the image 22
' ) will be found.

Now, it is possible for one person to determine that the sample 21 and the image 22 correspond and input it into a computer, or it is also possible to have the computer determine the correspondence. The latter case will be explained below with some additional information.

First, before measurement, the frequency of the RF pulse or the magnitude of the static magnetic field is adjusted so that the center of the phantom becomes the center of the reproduced image. Under these conditions, if an image as shown in Fig. 2 is obtained with N points on the horizontal axis and N points on the vertical axis, ■I=O, J'=0 (the upper left point, Starting from the point on the horizontal axis and the vertical axis, scanning is performed in the direction shown by the arrow in Figure 5 to detect the maximum value.However, since the image contains noise, it is smoothed in advance and the , it is necessary to set an appropriate threshold and remove background noise. Therefore, ■ Set the coordinates where the maximum value was obtained as (IP, J,), and set the values at all other points to zero. Next, as shown by the arrow in FIG. 6, scan again from the point I=J=0, and when a value other than zero is found, add 1 to J and scan again from I=O, repeating the process. At this time, if the following conditions exist between the coordinates of the horizontal axis of adjacent maximum values, scanning is stopped there,
The coordinates of the point that is the largest among the maximum values detected so far are stored in the memory 19 as Cxx, J).

l r, -1, , l > ΔI where, = o, 1. ..., N-2, and Δwork can be determined appropriately while looking at the results. Once (t, J1) is determined, (1) set all the maximum values detected up to that point to zero, return to step (2), repeat the steps up to (2), and stop scanning when all the values become zero.

Through such operations, the positions of all the samples on one image have been determined. In this way, (I', J
' ) is part of the image, so the remaining points must be found by interpolation. The sample position on the phantom can be determined in advance, so if it is (I, J), the coordinates (I', J') obtained by the above operation are
It can be obtained using equation (8) from . Exactly the same operation yields . From these four equations, ε, (I, J).

εF(I, J) and E(I, J) are found. Note that the method for determining the correspondence between the sample and the image is not limited to this, and for example, calculating the center of gravity may be considered.

In the description of the above embodiments, 1. Although the modified spin warp method was used, it is clear that the present invention is not limited to this and is effective for other sequences as well.

〔Effect of the invention〕

As described above, according to the present invention, in an inspection apparatus that utilizes NMR phenomena in a static magnetic field, a gradient magnetic field, and a high-frequency magnetic field, the intensity distribution within the field of view of the static magnetic field and/or the gradient magnetic field is calculated, and Since the image obtained under the magnetic field is corrected point by point, it is possible to realize a device that can correct the effects of non-uniformity of the static magnetic field and/or non-linearity of the gradient magnetic field. It is something that plays.

[Brief explanation of the drawing]

Figure 1 is a diagram showing a phantom for measuring the distribution of static magnetic fields and gradient magnetic fields, Figure 2 is a diagram showing images obtained from it, and Figure 3 is a diagram to explain the modified spin warp method. , FIG. 4 is a diagram showing a schematic configuration of an inspection apparatus which is an embodiment of the present invention, and FIGS. 5-6 are diagrams for explaining a method of determining the center position of a sample from an image. Figure 1 Figure 20

Claims (1)

[Claims] 1. Magnetic field generating means for a static magnetic field, a gradient magnetic field, and a high-frequency magnetic field, a signal detecting means for detecting a nuclear magnetic resonance signal from an object to be examined, and calculating a detection signal of the signal detecting means. In an inspection apparatus using nuclear magnetic field resonance, which includes a computer and an output means for the calculation results of the computer, and is configured to measure orthogonal coordinate points in Fourier space of the inspection target, the intensity of a gradient magnetic field applied during signal observation. A method using nuclear magnetic resonance, characterized in that it is configured to include a memory for storing images obtained by changing the image in a plurality of ways, and to read out the memory and calculate the intensity distribution of the static magnetic field and/or the gradient magnetic field. Inspection equipment. 2. In the inspection device according to claim 1,
An inspection apparatus using nuclear magnetic resonance, characterized in that the apparatus is configured to correct an image measured under the magnetic field using data read from a memory storing the magnetic field intensity distribution.